CN101887595A - Three-dimensional digital earth-space data organizing and rendering method based on quad-tree index - Google Patents

Abstract

The invention provides a three-dimensional digital earth-space data organizing and rendering method based on a quad-tree index, which belongs to the technical fields of cartography, geography information systems and virtual reality. The method comprises the following steps of: unifying common spatial data formats under multi-scale and multi-projection conversion, multispectral, multi-temporal and high-resolution remote sensing satellite images and aerial images as well as digital thematic maps with different scales into same coordinate system, carrying out operation on the attribute of each element and the parameter regulation of quad-tree tiles, outputting in the form of the quad-tree tiles, carrying out quad-tree cutting on three-dimensional landscape map data, leading spatial data into a relevant database, and carrying out unified management. By using the method, common vector data, raster data, altitude data and three-dimensional map data are organically integrated and issued into a three-dimensional digital earth prototype, thereby shortening the time of data preprocessing, improving the execution efficiency and providing a new integration method for three-dimensional digital earth fundamental geographic data dissemination.

Description

3-dimensional digital geospace data organization rendering intent based on the quaternary tree index

Technical field

The invention belongs to Geographic Information System, map cartography and virtual reality technology field, be specifically related to 3-dimensional digital geospace data organization rendering intent based on the quaternary tree index.

Background technology

Digital earth is based on computer technology, multimedia technology and extensive memory technology, is that tie uses the magnanimity Earth Information earth to be carried out the three-dimensional description of multiresolution, multiple dimensioned, multi-space and multiple class with the broadband network.

Digital earth mainly is made up of spatial data, text data, operating platform, application model.Spatial data mainly comprises all kinds of digital thematic map of all kinds of multispectral, the multidate of the spatial data of global large, medium and small engineer's scale and the earth, high-resolution remote sensing satellite image, aviation image, different proportion chi.A basic thought of digital earth is that high-resolution remote sensing image is covering the whole world, and by setting up digital elevation model covering the whole world and vector data, sets up earth landscape model, forms virtual earth.

Spatial data mainly is to realize by the mode that various dissimilar spatial datas combine.The surface area of the earth is near (1.5 * 10 ¹²) square kilometre, so the data volume of the 3-dimensional digital earth is huge especially and data type is complicated, data volume increases progressively along with the raising of resolution is geometric series, often reaches a hundreds of G, even tens TB bytes.Under the present network environment or on the PC, so huge data are difficult to reach the requirement of real-time Transmission and demonstration, so the real-time rendering of the transmission of magnanimity geodata and three-dimensional scenic becomes the important content in the digital earth research.The large-capacity data storage that drops to of the renewal of computer memory device and price provides support, concerning web-based system, rationalization's spatial data, improve the access speed of mass data and reduce the transmission of unnecessary data on network as much as possible, significant to the demonstration speed that improves global three-dimensional scenic.

Whole world terrain data amount is very huge, and for example, 1: 25 ten thousand the total data volume of terrain data in the whole world reaches more than 20 G, adds image, and its data volume is well imagined.In order to reduce the consumption of calculator memory, accelerate computer processing time and save network latency, the ultra-large geodata of visit multiresolution that can be the fastest, must design a rational database storing structure and be used for managing flexibly from coarse to fine data, such Adaptive Data library structure must be realized by the usage space index.Usually with the data hierarchy tissue, the data of each layer are divided into little data block again, and this method of data of organizing is commonly referred to quad-tree structure or pyramidal data structure.Fig. 1 has illustrated the layer of structure relation of quaternary tree.

List of references

[1] adopt as heir Li Qi, Yi Shanzhen. NSII and digital earth [M]. Beijing: publishing house of Tsing-Hua University, 1999.

[2] Gong Jianya, Du Daosheng, Li Qingquan, etc. contemporary geographical information technology [M]. Beijing: Science Press, 2004:87～88.

[3] side is abundant, Zhou Chenghu, and Jing Guifei, etc. the 4th generation GIS software study. Chinese image graphics journal, 2001b, 6A (9): 817～823.

[4] Song Guanfu, clock is pleasing to the ear. component geographic information system research and development, image graphics journal, 1998,4:314.31.

[5] Tong Xiaochong beautifully adornedly advances Zhang Yongsheng. the structure of global multiresolution data model with show fast. mapping science .2006.1.

[6] white for army building, State of Zhao wins, and Chen Jun is based on the global terrain visualization of the ellipsoid triangulation network, Wuhan University's journal, 2005.05.

[7] Li Deren. information highway, spatial data infrastructure and digital earth. mapping journal, 1999.

[8] Li Deren. the opinion century remote sensing and development. Wuhan University's journal, 2003.02.

[9] visual research and the some application of the blue .DEM of Wu Yan. Surveying ﹠ Cartography Scientific ﹠ Technological Univ., Wuhan's academic dissertation, 1998.

[10] Peng Hui. based on the foundation and the applied research thereof of the digital elevation model of MicroStation, Surveying ﹠ Cartography Scientific ﹠ Technological Univ., Wuhan's academic dissertation, 1998.

[11] Xue Yong, Wang Jianqin, Guo Huadong. the young view of digital earth grid computing [J]. remote sensing journal, 2004.8 (I).

[12] Yang Yongchong, Guo Dazhi. based on the Model of Digital Maps [J] of geographic coordinate, Southwest Jiaotong University's journal, in June, 2005, Vol 40, No3.

Summary of the invention

The objective of the invention is to propose a kind of 3-dimensional digital geospace data organization rendering intent based on the quaternary tree index.This method can be with global multi-scale, the Data Format commonly used under many projections change and the earth all kinds of multispectral, multidate, high-resolution remote sensing satellite image, aviation image, all kinds of digital thematic map of different proportion chi is unified in same coordinate system, and can carry out each attribute of an element is regulated, operations such as quaternary tree tile parameter regulation, visualized data is exported with quaternary tree pyramid tile form the most at last, the three-dimensional sight map datum is carried out the quaternary tree cutting, the geometric data and the attribute data of spatial data are imported in the relevant database unified management.The present invention detects framework for final data provides a cover, has guaranteed that data can be stablized, the correct Digital Earth Platform that offers.

Technical scheme of the present invention based on the 3-dimensional digital geospace data organization rendering intent of quaternary tree index, comprises following mentality of designing (perhaps following basic step):

1) adopts digital earth data hierarchy piecemeal mechanism

2) generate tile and it is played up

3) multi-source data is carried out fusion treatment

4) adopt 3-dimensional digital earth prototype data testing tool

Beneficial effect of the present invention is: vector data commonly used, raster data, altitude figures, three-dimensional map data etc. can be organically blended by using this method, and be published in the 3-dimensional digital earth prototype with the pyramidal cutting storage mode of quaternary tree, significantly shortened the pretreated time of data and improved execution efficient, for the issue of 3-dimensional digital earth geo-spatial data provides new integrated approach.

Description of drawings

The layer of structure relation of Fig. 1 quaternary tree

Fig. 2 cartesian coordinate system

Fig. 3 tile wide and high

Fig. 4 36 degree are divided

Fig. 5 different stage is divided synoptic diagram

Each rank quaternary tree tile piecemeal of Fig. 6 is represented

Fig. 7 n layer quaternary tree is to the rule of cutting apart of n+1 layer quaternary tree

Fig. 8 tile is preloaded into and the transition display process

Fig. 9 map datum source

Figure 10 is three-dimensional sight map render process in 3Ds Max

Figure 11 single width map cutting flow process

Several related cutting flow processs of Figure 12 super large map

The hierarchical structure relation of Figure 13 linear vectogram layer and DEM and image

Figure 14 supports terrestrial coordinate to arrive the Lambert projection coordinate

Figure 15 supports the Jpeg2000 raster data

Figure 16 supports dem data

The elementary factor data model of Figure 17 OpenGIS

Figure 18 tile rendering intent instrument prototype sectional drawing

Figure 19 tile rendering intent instrument prototype derived data is to 3-dimensional digital earth prototype

Figure 20 multilayer grid merges mutually with vector data

Embodiment

Introduce mentality of designing of the present invention below in conjunction with Figure of description.

1 digital earth data hierarchy piecemeal mechanism

1.1 face of land tile is divided mechanism

Take Cartesian coordinates, (X=0, Y=0.) in the projection coordinate lower left, promptly Geophysical South Pole (90 ,-180) (representing with longitude and latitude) is seen Fig. 2 and Fig. 3 to initial point.

As Fig. 3, (all tiles all are squares to use " Level Zero Tile Size " to decide the wide and high size of each tile.The level zero tile size of standard does not also formulate, but it must satisfy and can be divided exactly by 180.

With the earth 360 degree longitudes and 180 degree latitudes is standard, and ground floor is divided with 36 degree, sees Fig. 4.

Be divided into and be (360/36) * (180/36)=50 layer, the second layer is with 18 degree, analogizes in proper order with 9 degree for the 3rd layer, sees Fig. 5 and Fig. 6.

1.2 tile location mechanism

Definition:

A. the map grade for since 1, the ranks number is for several since 0;

B. the extreme value XMin of body of a map or chart, XMax, YMin, YMax represents;

C. represent Diff E with the x Δ, the y Δ is represented Diff N;

D. the mode that comprises map grade, ranks number is adopted in the name of quaternary tree map tile, and as " zoomx_row_column ", wherein x represents the map grade, and row and column analyze expression residing line number of current tile and columns;

1.2.1 map quantity is calculated

If first grade map is the capable n row of m, then the map ranks number of level level is as follows:

Line number: m * 2 ^Level-1, level 〉=1

Columns: n * 2 ^Level-1, level 〉=1

1.2.2 the conversion of geographic coordinate and placement of tiles

1.2.2.1 calculate geographic range by placement of tiles

The input: ranks to (i, j)

Output: the Extent scope (currentXMin, currentXMax, currentYMin, currentYMax)

The coordinate range of the capable j column position section of i is calculated by following formula under the level level:

$\mathrm{currentXMin}=\mathrm{XMin}+j\×\frac{\mathrm{\Δx}}{m\×2^{\mathrm{level}-1}}$

$\mathrm{currentXMax}=\mathrm{XMin}+(j+1)\×\frac{\mathrm{\Δx}}{m\×2^{\mathrm{level}-1}}$

$\mathrm{currentYMin}=\mathrm{YMax}-(i+1)\×\frac{\mathrm{\Δy}}{n\×2^{\mathrm{level}-1}}$

$\mathrm{currentYMax}=\mathrm{YMax}-i\×\frac{\mathrm{\Δy}}{n\×2^{\mathrm{level}-1}}$

1.2.2.2 by the coordinate Calculation placement of tiles

Input: (x, y) coordinate

Output: i, j (the ranks number of picture)

Wherein,

$j=\sqrt{(x-X_{\mathrm{Min}})/\mathrm{\Δx}\×m\×2^{\mathrm{level}-1}}$

$i=\sqrt{(Y_{\mathrm{Max}}-y)/\mathrm{\Δy}\×n\×2^{\mathrm{level}-1}}$

Adopt the data structure of quaternary tree, with reference to Fig. 4, Level 0 is divided into 50 zones to the earth by 36 degree, and each block splitting of the 0th layer is four subtrees, and this one deck is 1 layer of Level, calculates as Fig. 7 successively.

Each piece all has four subtrees, comprises self-information, and data definition is:

Double West; The plane Cartesian coordinates value of // four direction

double?East；

double?North；

double?South；

Angle CenterLatitude; // tile central authorities longitude and latitude

Angle?CenterLongitude；

double?LatitudeSpan；

double?LongitudeSpan；

Int Level; // level Level

Int Row; // OK, row

Int Col; // subtree

QuadTile?northWestChild；

QuadTile?southWestChild；

QuadTile?northEastChild；

QuadTile?southEastChild；

One 512 * 512 the texture of the corresponding current Level of each piece, every limit of each piece is divided into 40 equal portions, generates 41 * 41 or 43 * 43 summits, again in conjunction with the elevation information real-time rendering.

Because the structure of hierarchical block, during to the loading of terrain data and real-time rendering, only need be from the downloaded need hierarchy, texture picture, the elevation information in zone, can also carry out the direct pinup picture of picture that pre-service becomes three-dimensional fruit to some picture, not need elevation information.

Transition processing:

Level (n+1) level has not been downloaded not treatable picture, still shows the effect of Level n level, sees as Fig. 8.

2 generate tile

2.1 three-dimensional tile data

The pre-generation of map is the square picture that the map of specified scope is cut into several rows and row according to specified size and specified format (as TIFF, JPEG, PNG etc.), cut the map section that figure obtains and be called tile (Tile), tile obtain flow process as shown in Figure 9.Play up the grid map that tile can only obtain two-dimensional vector data, image data or digital elevation data correspondence, the cutting figure and need carry out three-dimensional modeling of three-dimensional sight map by instruments such as AutoCAD, 3DMax, and play up the view picture three-dimensional map again after regulating by the VRay plug-in unit.

Three-dimensional actual in the electronic three-dimensional map is 2.5 abundant dimensions of a kind of sightingpiston, is the fixed viewpoint static rendering picture that is generated by three-dimensional model.To determine camera angle according to regional orientation of room characteristics before making, make software set camera, lighting effects, making base map, form the standard scene, use for the producer is unified at three-dimensional model.Those are blocked, cannot see or be positioned at the object of backseat, can omit, simplify.Point, line, surface city entity information figure layer in the three-dimensional scenic, layering is made when modelling, and the three layer model stack forms the city three-dimensional sight.Figure 10 shows that in 3DSMax three-dimensional model plays up the implementation procedure of view picture.

2.2 tile cutting

Because the restriction of computer hardware, has restriction at technology realization aspect .Net class libraries for image volume, map length and width amplitude for the single width that is written into is limited in the 15000*15000, and therefore the cutting with map is divided into the cutting of single width map and several related cutting two classes of super large map:

1, single width map cutting flow process:

Step 1:, make its value of growing tall be the multiple of tile to the original value of cut coverage is revised;

Step 2: (northwest) begins this fixed range is cut figure from the upper left corner, adopt the quaternary tree data structure between the follow-up different brackets map, tile (Tile) on the level level will be fissioned into 4 to the level+1 level, this structure helps Qie Tu and demonstration, but the map that draws does not have fixing engineer's scale, therefore engineer's scale changes with geographical ordinate, is not according to engineer's scale when carrying out the calculation of geography amount but directly calculates according to geographic coordinate.The map naming rule is N level map picture called after zoomN.png.

The tile naming rule of the capable chunkcol of x level chunkrow row is zoomx_chunkrow_chunkcol.png, the naming rule of cutting back tile be zoomx_chunkrow_chunkcol x_chunkrow_chunkcol.

Each variable-definition wherein:

X: current level of zoom;

Chunkrow: the line number of current tile in map;

Chunkcol: the columns of current tile in map;

2, several related cutting flow processs of super large map:

Step 1: the super large map is divided into some little maps that are mutually related, little map naming rule be uperchunk_row_col zoomN.png, wherein row is a shared line position in the little map of current little map after all cuttings, and col is a shared column position in the little map of current little map after all cuttings.

Step 2: adopt single width map cutting flow process to cut to little map

The naming rule of cutting back tile be superchunk_row_col|zoomx_chunkrow_chunkcol x_chunkrow_chunkcol.

Each variable-definition wherein:

Row: shared line position in the little map of current little map after all cuttings;

Col: shared column position in the little map of current little map after all cuttings;

X: current level of zoom;

Chunkrow: the line number of current tile in map;

Chunkcol: the columns of current tile in map;

Several related cutting flow processs of single width map cutting flow process/super large map are as Figure 11 and Figure 12.

3 multi-source datas merge

3.1 fused data figure layer

Along with the difference of viewpoint distance, the sharpness of dimensional topography changes in real time, and at this moment the system's DEM and image data of calling in different resolution automatically constructed terrain model.When landform from a level of detail during to another level of detail transition, system calls several terrain datas simultaneously and shows that cutting apart of landform dough sheet will change.If atural object does not merge mutually with landform, will cause the space vector entity " sinking " and " raising " phenomenon to occur.In order to accelerate speed of displaying and graphical quality, the data of polar plot layer are the same with DEM and image wants differentiated control, as Figure 13.

3.2 projection support conversion

By reading the corresponding .prj file of .shp file, and resolve, as Figure 14 with the WKT format standard.

Parameter comprises:

Coordinate system is to describe by the mode of OpenGIS WKT (Well known Text).It comprises following information:

A. the title of a total coordinate system.

B. the title of a geographical co-ordinate system.

C. the earth reference system.

D. reference ellipsoid, the ellipsoid semi-major axis and.The derivative of flattening of ellipsoid (being a/ (a-b)).

E. the name car and it is in the departure degree of the first meridian of one 0 degree warp.

F. projection type.

G. the tabulation of a projective parameter.

H. an organization and be converted into the rice or the transforming factor of radian.

I. Zhou title and ordering.

3.3 many kinds of vectors and raster data type are supported

As unified interface, can support different pieces of information with the OpenGIS standard on the same stage, as

■ vector data: Shape data, mapinfo data, gml data, geojson data

■ raster data: Geotiff data, ecw data, jp2 data (seeing Figure 15)

■ digital elevation data: dem data (seeing Figure 16)

■ gps data: .GPX form

Use GDAL has realized in the storehouse support to Shape data, mapinfo data, gml data, geojson data isovector data.

Vector data is based on rectangular coordinate system, describes geographic element with point, line, polygon and gets data model or data structure.Each geographic element has a series of sequential coordinates, and (these key elements combine with attribute for x, y) coordinate description.We have designed the strict Geometry model of following the elementary factor data standard of OGC OpenGIS, and what define among the Geometry is basic geometric configuration in the vector graphics, and some simple space correlation relationship analysis functions.Most important geometric configuration is point, line, surface, and the multiple spot that derives from thus, multi-thread, multiaspect, curve, straight line, loop wire, polygon etc., sees Figure 17.

The relation of grid positions (pixel coordinates or line coordinates) and Geographic Reference coordinate is described in two ways.First kind also is that a kind of of widespread usage is affined transformation.

Affined transformation is made of six coefficients, and it is mapped to the Geographic Reference space with following relational expression with pixel/line coordinates:

Xgeo＝GT(0)+Xpixel*GT(1)+Yline*GT(2)

Ygeo＝GT(3)+Xpixel*GT(4)+Yline*GT(5)

For the image of downward arrangement, GT (2) and GT (4) are zero, and GT (1) is that pixel is wide, and GT (5) is the pixel height, the upper left corner coordinate of (GT (0), GT (3)) expression grid top left pixel.

Attention: top pixel and line coordinates all are to arrive the lower right corner coordinate (pixel is wide, the pixel height) of lower right corner pixel from the upper left corner coordinate (0.0,0.0) of top left pixel.The center of top left pixel/line then is (0.5,0.5).

Definition: wave band

GDALRasterBand class among the corresponding GDAL of the raster data of a wave band.It has described the band/channel/layer of single wave band.It is not once to describe whole video.For example, the RGB video of a 24bit generally is described to three wave bands in a dataset, respectively color among the corresponding red/green/blue three.

Each wave band is explained and is specially:

GCI_Undefined: acquiescence, unknown message.

GCI_GrayIndex: be an independently gray scale video.

GCI_PaletteIndex：this?raster?acts?as?an?index?into?a?color?table。

The red wave band of GCI_RedBand:RGB or RGBA video.

The green wave band of GCI_GreenBand:RGB or RGBA video.

The blue wave band of GCI_BlueBand:RGB or RGBA video.

The alpha passage of GCI_AlphaBand:RGBA video.

The tone of GCI_HueBand:HLS video.

The saturation degree of GCI_SaturationBand:HLS video.

The light intensity of GCI_LightnessBand:HLS video.

The cyan wave band of GCI_CyanBand:CMY or CMYK video.

The magenta wave band of GCI_MagentaBand:CMY or CMYK video.

The yellow wave band of GCI_YellowBand:CMY or CMYK video.

The black wave band of GCI_BlackBand:CMYK video.

The support data attribute is regulated: support data attribute to regulate

3.4 quaternary tree tile parameter regulation

Tile is derived after by the quaternary tree parameter regulation, and the derived parameter configuration file, form as:

<MapParam>

<tilesize>128</tilesize>

<ColumnNum>4</ColumnNum>

<RowNum>2</RowNum>

<centerX>-243383.913445131</centerX>

<centerY>4376213.31793891</centerY>

<ZoomParam>2</ZoomParam>

<MaxLevel>4</MaxLevel>

<Xmin>-4663419.25215648</Xmin>

<Xmax>4176651.42526622</Xmax>

<Ymin>2166195.64858323</Ymin>

<Ymax>6586230.98729458</Ymax>

</MapParam>

3.5 the foundation of spatial database

A. database connects

Comprise: 1. the Sql mode is verified; 2. Windows Authentication.

B. the newdata storehouse comprises database file (.MDF) and journal file (.LDF).

C. spatial database is disposed: generate field " BLUE.GEOMETRY_COLUMNS " and " BLUE.SPATIAL_REF_SYS " in specified database, be respectively applied for geometric graph layer and the space W KT information deposited.

D. vector data imports

With the shape data importing in database.The shp file generates fixed field oid, the_geom, zID, the_geom_Envelope_MinX, the_geom_Envelope_MinY, the_geom_Envelope_MaxX, the_geom_Envelope_MaxY deposits serial number, geometric properties data, figure level number, bounding box horizontal ordinate minimum value, bounding box ordinate minimum value, bounding box horizontal ordinate maximal value, bounding box ordinate maximal value respectively.Generate the attribute information field by the .dbf file, for example: " city ", " population ".

3.6 geodata is searched in batches

Longitude and latitude search based on Geocoding searches corresponding longitude and latitude by the input geographic position.Send form and be " Http:// ditu.google.com/maps/geo? output=xml﹠amp; q=geographic position " request, the xml formatted file that returns is resolved, point.coordinates data wherein are latitude and longitude value.By using data set and data adapter unit, carry out data search in batches based on sqlserver2005.

4 by this method map is edited and derived data to the demonstration of digital earth, see Figure 18, Figure 19 and Figure 20.

Claims (2)

1. based on the 3-dimensional digital geospace data organization rendering intent of quaternary tree index, comprise following mentality of designing:

(1) adopts based on quaternary tree pyramidal digital earth data hierarchy piecemeal and tile location mechanism, geographic coordinate and tile are carried out mutual computing, calculate geographic range by placement of tiles;

(2) design based on the pyramidal three-dimensional map of quaternary tree play up with the single width map, on a large scale several the associating maps the tile generating algorithm;

(3) design is supported the general elementary factor model of vector data support and is supported the general wave band of all kinds of raster datas, altitude figures to read structure, and the method unified coordinate system that multiple projective transformation is converted to terrestrial coordinate is united, and multi-source data is carried out fusion treatment.

2. the 3-dimensional digital geospace data organization rendering intent based on the quaternary tree index according to claim 1, it is characterized in that, in the described step (1), at the 0th grade the earth is divided into 50 zones by 36 degree, one 512 * 512 the texture of each corresponding current Level in zone, every limit of each piece is divided into 40 equal portions, generates 41 * 41 or 43 * 43 summits;

In the described step (2), three-dimensional model layering stack forms the city three-dimensional sight, and the three-dimensional map of playing up is cut into single width map, several map tile of super large map;

In the described step (3), set up two-dimentional cartesian coordinate system and support multi-source data to merge, store fusion for point, line, surface, multiple spot, multi-thread, multiaspect, curve, straight line, loop wire, polygon with vector data is abstract; Then, raster data is carried out affined transformation according to coordinate system, and extract the data fusion of three wave bands of red/green/blue; At last, the bounding box information and the attribute information of element in the fused data is saved in the relevant database, takes Geocoding to carry out index.

Images